Communication Networks and Protocols
This domain examines how data is encapsulated, addressed, routed, and delivered across LANs and WANs. Expect questions on layered models, Ethernet switching, IPv4/IPv6 addressing and subnetting, routing, transport protocols, and core Internet services.
Core rules
- Encapsulation adds headers at each layer; decapsulation reverses the process at the receiver. The TCP/IP model maps loosely to the OSI seven layers.
- Ethernet switching learns source MAC addresses and forwards frames by destination MAC; unknown unicast is flooded. VLANs use IEEE 802.1Q tags.
- IPv4 addressing: a /n prefix leaves 32 − n host bits. Usable hosts = 2^(32−n) − 2, excluding network and broadcast addresses. CIDR allows longest-prefix routing.
- ARP resolves IPv4 to MAC inside a broadcast domain; IPv6 uses Neighbor Discovery instead.
- TCP provides connection setup with a three-way handshake, sequence numbers, acknowledgements, retransmission, flow control, and congestion control (slow start, congestion avoidance, fast retransmit). UDP is connectionless and has no delivery guarantee.
- DNS resolves names to records; DHCP assigns addresses, masks, gateways, and DNS servers.
Worked example
Is 192.168.10.130 usable in 192.168.10.0/26? A /26 mask is 255.255.255.192, with block size 64. Subnets start at .0, .64, .128, .192. The .128/26 subnet covers .128–.191: .128 is the network address, .191 is broadcast, and .129–.190 are usable. Therefore .130 is usable, with broadcast .191.
Traps
- Forgetting to subtract two addresses for network and broadcast.
- Assuming a /26 is a /24 with fewer hosts rather than recalculating the block boundary.
- Confusing TCP port numbers with IP addresses.
- Claiming UDP is reliable because applications add their own acknowledgements.
- Treating MAC addresses as routable across the Internet.